Disclosed are photonic particles and methods of using particles in biological samples. The particles are configured to emit laser light when energetically stimulated by, e.g., a pump source. The particles may include a gain medium with inorganic materials, an optical cavity with high refractive index, and a coating with organic materials. The particles may be smaller than 3 microns along their longest axes. The particles may attach to each other to form, e.g., doublets and triplets. The particles may be injection-locked by coupling an injection beam into a particle while pumping so that an injection seed is amplified to develop into laser oscillation. A microscopy system may include a pump source, beam scanner, spectrometer with resolution of less than 1 nanometer and acquisition rate of more than 1 kilohertz, and spectral analyzer configured to distinguish spectral peaks of laser output from broadband background.
|
7. A flow cytometry system comprising:
a pump light source;
a flow channel;
a spectrometer; and
a spectral analyzer configured to distinguish spectral peaks of laser output from broadband background due to fluorescence.
13. A system comprising:
a pump light source configured to direct light comprising a predetermined wavelength toward a sample; and
a detector configured to measure a characteristic of an emission from a laser particle associated with the sample.
1. A microscopy system comprising:
a pump light source;
a beam scanner;
a spectrometer including:
a resolution of less than 1 nanometer; and
a spectral acquisition rate of more than 1 kilohertz; and
a spectral analyzer configured to distinguish spectral peaks of laser output from broadband background due to fluorescence.
2. The microscopy system of
3. The microscopy system of
4. The microscopy system of
6. The microscopy system of
8. The flow cytometry system of
9. The flow cytometry system of
10. The flow cytometry system of
11. The flow cytometry system of
12. The flow cytometry system of
14. The system of
a gain medium including one or more inorganic materials;
an optical cavity situated about the gain medium, the optical cavity having a refractive index greater than 2; and
a coating covering at least part of the optical cavity.
15. The system of
wherein the system further comprises:
a spectral analyzer configured to distinguish spectral peaks of laser output from broadband background due to fluorescence.
16. The system of
wherein the spectrometer is configured to have:
a resolution of 1 nanometer or less, and
an acquisition rate of more than 1 kilohertz, and
wherein the system further comprises:
a beam scanner.
18. The system of
20. The system of
21. The system of
22. The system of
an arrangement for fluorescence microscopy comprising:
the second light source; and
the second detector.
23. The system of
an arrangement for fluorescence cytometry comprising:
the second light source; and
the second detector.
24. The system of
wherein the detector comprises an interferometer configured to measure fringes between an injection beam and output from laser particles, and
wherein the system further comprises:
an injection laser configured to control laser particles by injection locking;
a receptacle for receiving a sample containing laser particles.
25. The microscopy system of
receive, over a period of time, a plurality of spectral outputs from the spectrometer, each spectral output indicative of an intensity detected by the spectrometer at each of a plurality of wavelengths;
determine a lifetime of each of a plurality of spectral peaks in the plurality of spectral outputs; and
distinguish one or more spectral peaks due to narrowband emissions from the one or more laser particles from fluorescence based on the respective lifetime of each of the plurality of spectral peaks.
26. The flow cytometry system of
receive, over a period of time, a plurality of spectral outputs from the spectrometer, each spectral output indicative of an intensity detected by the spectrometer at each of a plurality of wavelengths;
determine a lifetime of each of a plurality of spectral peaks in the plurality of spectral outputs; and
distinguish spectral peaks due to narrowband emissions from the laser particles from fluorescence amongst the plurality of spectral peaks based on the respective lifetime of each of the plurality of spectral peaks.
27. The microscopy system of
identify a plurality of wavelengths associated with the spectral peaks of laser output; and
determine a characteristic associated with each of the laser particles contained in the sample based on the plurality wavelengths.
28. The microscopy system of
29. The flow cytometry system of
identify a wavelength associated with a spectral peak of laser output; and
determine a characteristic associated with a laser particle contained in the sample based on the wavelength, wherein the data comprises characteristics associated with the laser particles contained in the sample.
30. The flow cytometry system of
a sorting apparatus configured to sort cells associated with the laser particles based on the wavelengths of spectral peaks identified in connection with the cells.
|
This application is a continuation of U.S. application Ser. No. 16/306,278 filed Nov. 30, 2018 which is a U.S. National Phase application of PCT/US2017/035923 filed Jun. 5, 2017 which is based on, claims priority to, and incorporates herein by reference in its entirety U.S. Provisional Application Ser. No. 62/345,070, filed Jun. 3, 2016, and entitled, “Laser Micro-Particles.” The references cited in the above provisional patent application are also hereby incorporated by reference.
This invention was made with government support under ECCS-1505569 awarded by the National Science Foundation, and under DP1-DB024242 awarded by the National Institutes of Health. The government has certain rights in the invention.
This document concerns an invention relating generally to miniature lasers that can be embedded, implanted, or injected into samples such as biological cells and tissues, and more specifically, to optically excitable laser particles made of organic and inorganic materials, and the fabrication, functionalization, delivery, and imaging of the laser particles, and their use as probes for massively parallel imaging, sensors, and assays.
Fluorescent probes, such as dyes, fluorescent proteins and quantum dots, have become indispensable tools in biomedical imaging, cell sorting, immuno-histology, high-throughput screening, and numerous other biochemical measurements. Although these luminescent probes are immensely useful, their relatively broad emission spectra, typically 30-100 nm, limit the number of probes that can be simultaneously used without ambiguity and often make their spectra indistinguishable from the background emission of endogenous molecules in tissues. Conventional fluorescence microscopes are equipped to resolve 3 to 4 dyes, and state-of-the-art cytometry is limited to eleven channels. Multiplexing four different dyes can give 16 (=24) combinations. Simultaneous expression of three genes encoding blue, green, and red fluorescent proteins at different ratios in cells, as in Brainbow and RGB marking, can generate hundreds of colors. However, the transfection is stochastic, and the fidelity of color reading is prone to noise. To date, the number of fluorescence colors for imaging has been limited to less than a dozen.
It is fundamentally challenging to engineer fluorophores for much narrower emission linewidth because of the quantum-mechanical broadening of the electronic levels in molecules. The irregular shapes and thermodynamic fluctuations resulted in spectral broadening of emission from semiconductor quantum dots. The attenuation of plasmonic electron oscillations in metallic nanoparticles resulted in emission widths of >50-100 nm. By comparison to these electronic resonance, optical resonance offers effective approaches to generate narrow emission lines. A laser is a great example. By placing fluorophores and semiconductor materials inside an optical cavity, an extremely narrow spectral line can be produced. The output of a single-frequency laser can be a millionth of nanometer in wavelength, tunable over the entire gain width by changing the cavity resonance.
The present disclosure provides example photonic particles comprising at least one gain medium and at least one optical cavity, and having their largest dimension not substantially greater than 3 μm in preferred implementations, or no greater than 2 μm in other preferred implementations. The gain medium contains a sufficient number of gain elements, such as fluorophores and electron-hole pairs, and the cavity has a sufficiently low optical loss so that when the gain medium is excited or stimulated by pump light at sufficiently strong intensity levels, the gain elements emit light that exhibit spectral characteristics defined by the optical resonance modes of the cavity. The particles may be made of semiconductor materials with appropriate shapes and structures, such as quantum well structures, spherical shapes, or multiplayer Bragg reflectors. Alternatively, the particles may be configured with fluorescent gain molecules, high reflective-index dielectric resonators, and potentially metals.
Example laser particles have an output emission spectra with one or plural peaks with each linewidth narrower than 5 nm in preferred implementations, and typically less than 1 nm or even 0.3 nm in other preferred implementations. The peaks are primarily determined by the resonance of the optical cavity in the particles. The center wavelength of the output spectra may cover the entire visible and near-IR range, for example, from 400 to 1900 nm.
In one aspect, the present disclosure provides an optical system comprising at least one pump light source to excite the laser particles and at least one detection arrangement. The detection arrangement comprises spectral resolving elements, such as diffraction granting and dichroic filters. The pump source includes pulsed lasers with pulse widths in the range from 100 fs to 10 ns.
A specific preferred embodiment for laser particles includes quantum-well micro disk lasers, standalone surface emitting Bragg reflector semiconductor lasers, semiconductor spheres, all with diameters or lengths in the range of 500 nm to 3 μm.
The foregoing and other advantages of the disclosure will appear from the following description. In the description, reference is made to the accompanying drawings, which form a part hereof, and in which there is shown by way of illustration a preferred embodiment of the disclosure. Such embodiment does not necessarily represent the full scope of the disclosure, however, and reference is made therefore to the claims and herein for interpreting the scope of the disclosure.
The foregoing and other aspects and advantages of the present disclosure will appear from the following description. In the description, reference is made to the accompanying drawings that form a part hereof, and in which there is shown by way of illustration one or more exemplary versions. These versions do not necessarily represent the full scope of the invention.
The present disclosure will hereafter be described with reference to the accompanying drawings, wherein like reference numerals denote like elements.
Before the present invention is described in further detail, it is to be understood that the invention is not limited to the particular embodiments described. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to be limiting. The scope of the present invention will be limited only by the claims. As used herein, the singular forms “a”, “an”, and “the” include plural embodiments unless the context clearly dictates otherwise.
It should be apparent to those skilled in the art that many additional modifications beside those already described are possible without departing from the inventive concepts. In interpreting this disclosure, all terms should be interpreted in the broadest possible manner consistent with the context. Variations of the term “comprising” should be interpreted as referring to elements, components, or steps in a non-exclusive manner, so the referenced elements, components, or steps may be combined with other elements, components, or steps that are not expressly referenced. Embodiments referenced as “comprising” certain elements are also contemplated as “consisting essentially of” and “consisting of” those elements.
The standard paradigm of using light to characterize or manipulate samples in biomedical applications, as depicted in
Lasers that are biocompatible and miniaturized to the size of a cell (or smaller) can be used inside a living system by being implanted in the body for certain duration, or injected into tissues and remotely operated. This new paradigm involves delivering light sources physically to the target. Referring to
A laser is comprised of three elements: a gain medium, a cavity, and pump energy source. Optical amplification in the gain medium is essential to generate stimulated emission. An amplifier with sufficient gain and long propagation length can generate amplified spontaneous emission. Although this process can generate output with laser-like characteristics, such as spectral narrowing, a narrow definition of laser requires optical feedback by a cavity. The cavity confines and makes resonance of light.
In addition to optical or photonic cavities, other type of cavities for quasiparticles, such as polaritons and plasmons, may be used to enable lasers. Polariton lasers based on exciton polaritons is a representative example of “non-photonic” lasers. The mechanism of polariton lasers is typically described in the context of the Bose-Einstein condensation of polariton. Surface plasmon amplification by stimulated emission of radiation or “spaser” commonly involves a metallic cavity for free electrons or plasmonic resonance.
Consider a gain medium containing active gain elements, such as dye molecules in aqueous solution or semiconductor atoms assembled in a solid-state crystal.
Simplified rate equations describing a laser particle in this type of four-level gain medium can be written as:
Here, N1(t) is the number of excited fluorophores in the gain medium, q(t) the number of photons in laser cavity modes, P(t) the pump rate, τs the spontaneous emission (fluorescence) lifetime of the gain molecules, and β the spontaneous emission factor that describes the fraction of spontaneous emission captured by laser modes. 1/τs corresponds to the spontaneous emission rate for all possible radiation modes in all frequencies and directions; thus, β/τs represents the rate of spontaneous emission coupled into the laser modes. The stimulated transition rate is proportional to the number of cavity modes, and equal to the spontaneous transition rate of the cavity modes multiplied by the number of photons: βN1(t)q(t)/τs.
For optical pumping, the pump rate can be expressed as:
where QY is the quantum yield that specifies the probability of one absorbed photon to generate one photon via either spontaneous or stimulated emission, σa is the absorption cross-section of the gain elements, and Ng(t) is the number of gain elements in the ground state. In most cases including the situation in
In Eq. (3), the intensity of pump light was assumed to be space invariant. This assumption is valid for laser particles with sizes much smaller than the size of the pump beam. For larger lasers with non-uniform pump, the pump rate is not solely determined by the peak pump power but also an overlap between the pump intensity profile and the laser mode profiles.
The output emission rate of the cavity modes Pl(t) and spontaneous fluorescence emission Pf(t) are given by:
where ql(t) and qf(t) are the numbers of emitted photons of the laser modes and fluorescence emission, respectively. It follows that at P(t)=Pl(t)+Pf(t), which states the conservation of the number of photons (given by the definition of P(t)).
Steady state (t>>τs): When the laser is in the steady-state state, (i.e. dN1/dt=dql/dt=0), we obtain
where
Pth=(βτc)−1. (8)
Laser oscillation enhances the “recycling” rate of the gain elements. Below the laser threshold, the number of excited elements N1 increases linearly with the pump rate, unless gain saturation occurs; i.e. N1 becomes comparable to Ntot (
Transient Build-Up State:
The governing equations, (1) and (2), describe the dynamics of laser oscillation. A necessary condition for the build-up of laser modes is dq(t)/dt>0 (at t=0; q=0), from which the minimum number of gain elements in the excited state should satisfy:
N1(t)>τs/βτc (9)
Consider situations where the pump energy is provided for duration of τp with a uniform rate. For q=0 (below threshold) and the square-profile pump pulses, the solution of Eq. (1) is:
N1(t=0)=Pτs[1−exp(−τp/τs)] (10)
The pump rate required to reach the threshold condition in Eq. (9) is:
For the case of quasi-continuous pumping (τp>τs), we confirm Eq. (8): Pth=(βτc)−1; the pump power at the lasing threshold is independent of the pulse duration, and the threshold pulse energy is proportional to the pulse duration. On the other hand, for the case of short-pulse pumping (τp<τs), we find Pth=(βτc)−1·τs/τp; the pump power increases as the pulse width decreases, and the threshold pump energy, Pthτp, is independent of the pump pulse duration.
From Eq. (9), the minimum number of excited gain elements in the laser should be greater than τs/(βτc). To have this many elements in the gain medium with a volume V, the absolute minimum concentration is greater than τs/(βτcV). For example, with τs=3 ns, τc=300 fs, and β=0.05, the gain medium should have at least 2×105 gain elements. For V=10 μm2, the theoretical minimum concentration is 330 nM, and V=1 μm2, a much higher concentration greater than 330 μM is necessary. In practice, most lasers require Ntot much greater than N1 because of the finite absorption of pump energy (further discussed below). Therefore, the minimum concentration of gain elements is considerably higher, by one or two orders of magnitude, than is estimated above.
Semiconductor and solid-state lasers are more compact and easier to maintain than dye lasers. However, for biomedical laser particles, fluorescent dyes are important, viable choices as gain materials. A variety of fluorescent dyes, with improved brightness and photo-stability and optimized bio-sensitivity over earlier laser dyes, may be used for fluorescence-based bio-imaging and biochemical assays.
The peak extinction coefficient (ε) of typical dye molecules ranges 10,000-100,000 M−1 cm−1. For example, let us consider molecules with ε=56,000 M−1 cm−1 and a fluorescence QY of 0.6. If a half of the molecules in the gain medium were excited to the upper level at a given time, we could obtain stimulated emission (gain) of up to 16,800 M−1 cm−1 (neglecting signal-induced gain depletion). A solution of these molecules at a concentration of 1 μM inside a 1-cm-long cuvette can produce a single-pass optical gain of 100.0168=1.039; that is, the optical intensity is amplified by 3.9% or 0.168 dB. The same amount of single-pass gain is achieved by a single mammalian cell containing the molecules at a concentration of 1 mM in the cytoplasm with a diameter of 10 μm.
Fluorescent proteins are proteins capable emitting fluorescence. The green fluorescent protein (GFP) is the first type found in the jellyfish, Aequorea victoria. The increasing popularity of fluorescent proteins in biomedical sciences has led to the continuing discovery of new wild-type proteins from various organisms and to the development of mutant variants with improved fluorescence characteristics. Several efficient fluorescent proteins have been derived from the jellyfish (eGFP, mCFP, eYFP, etc.), coral reef (DsRed, tdTomoto), and bubble-tip anemone (TurboRFP). FPs have high QY and large absorption and emission cross-sections, comparable to or even better than typical fluorescent dyes, and may be well-suited for use as the gain-medium of biological lasers. Compared to small-molecule organic dyes, the most unique, significant advantage of FPs is that they can be genetically encoded. This allows FPs to be expressed in specific target cell types by genetic targeting. This property of FPs may be used to achieve genetic specificity for the applications of laser particles based on FPs as gain materials.
Referring to
Inorganic semiconductors may be used as gain materials for lasers. Semiconductors are crystalline or amorphous solids with filled valence bands and empty conduction bands. Semiconductors differ from insulators in that the band energy gap, Eg, is typically less than 4 eV so that some electrons can transition to the empty conduction band by optical or thermal excitation, and the materials can be doped with impurities to alter their electronic properties. Most known semiconductor materials are crystalline inorganic solids. They include group IV materials, such as silicon (Si) and germanium (Ge), III-V compounds, such as GaAs and InP, and their alloys such as AlGaAs and InGaAsP, and II-VI compounds, such as CdS and ZnO. Oxides, such as TiO2, and 2-dimensional two-dimensional (2D) transition metal dichalcogenides, such as MoS2 and as WS2, are also semiconductors. Since laser action relies on efficient radiative processes, indirect-gap materials like silicon and germanium may not be suitable, and direct-gap III-V and II-VI compound semiconductors are preferred.
With respect to the physics of laser action in these systems,
To achieve a finite gain, population inversion is required, although schemes for laser without inversion exist. The number of electronic states in the conduction and valence bands is readily calculated from density of states of the carriers, ρc, which is:
for bulk (3D) semiconductor, where m* is the effective mass of the carrier. The number of electrons, N1, to fill up the conduction band of bulk semiconductor with volume V, up to energy ΔE above Ec is given by:
For example, for bulk GaAs the effective electron mass is me*=0.61×10−31 kg; at T=0 K and ΔE=0.1 eV, we get N1/V=2.5×1018 electrons/cm3, which is 4.15 mM in terms of an electron concentration, comparable to the highest density of organic fluorophores without quenching (see
Semiconductors efficiently absorb light at all wavelengths shorter than their band edge. The absorption coefficients (α, natural log basis) are ˜104 cm−1 in the optical frequency just above the band gap. The radiative lifetime τs of emission transition (electron-hole combination) is proportional to the population density of holes and the radiative recombination coefficient B (B≈10−10 cm3 s−1): approximately,
τs=(BN1/V)−1. (14)
For typical population density at lasing (N1/V≈1018-1019 cm−3), the radiative lifetime is 1-10 ns, same order of magnitude as fluorescence lifetimes of dyes.
Earlier semiconductor lasers exploited population inversion obtained at the interfacial region of a p-n homo-junction. Due to diffusion of minority carriers, the actual active region is relatively thick (˜1 μm), requiring high threshold pump energy. A more efficient design is a double hetero-structure, where the active region is delimited as a thin layer (˜100 nm) of a different semiconductor with lower band gap and higher refractive index. With this architecture, charge carriers are confined in the thin active layer, increasing population density. Higher N1 can be achieved by further confining the charges to a dimension of <<100 nm at which quantum effects become predominant, such as quantum wells (QW) with confinement in one dimension, quantum wires with 2D confinements, and quantum dots (QD) in 3D confinements. The density of states for QW is modified to ρc (E−Ec)=m*/πℏ2.
The gain spectral range of semiconductors can be tuned to some degree by varying adjusting the stoichiometry of the compound and inducing mechanical stresses. The emission spectra of 720-850 nm for AlGaAs and 900-1100 nm for InGaAs are relevant to biomedical applications as they match well with the optical window for light penetration in tissues. For emission in the visible, the InGaAlP system works in the red (630-700 nm), while for the blue-green region (400-530 nm) group-III nitrides (like GaN and InGaN) are commonly used.
The semiconductor materials may be grown via epitaxial processes on wafers from which the final devices are fabricated via lithographic processes. Nanoscale inorganic semiconductors may also be grown in the form of colloidal nanocrystals by chemical synthesis in solution. Various colloidal nanocrystals are extensively used in biological applications, mainly as fluorescent tags for bioimaging and biosensing. Due to the high quantum confinement in nanoscale dimensions, the band gap and thus the emission wavelength can be easily tuned by controlling the size of the particle. Typical nanocrystals are in the shape of spherical core/shell quantum dots, in which the core is the emissive material, while the shell usually serves as a passivation layer to reduce non-radiative recombination from surface defects. Moreover, the core/shell architecture has been shown to be important to obtain optical gain with these materials. More complex shapes have been also developed, like rods or branched nanocrystals, with advantages of increased lifetimes of radiative transition, controlled polarization and multiple emission bands.
Organic semiconductors are formed in condensed states incorporating fluorescent small molecules, oligomers or polymers, at a very high concentration so that the individual elements are coupled and collectively form a π-conjugated system with delocalized electrons. The π-conjugated system confers them strong electronic transition in the visible and the capability of conducting charges. The typical structure of an organic semiconductor is composed of a central conjugated backbone that determines the main optoelectronic properties, decorated with side groups that can be tailored to confer particular functionalities to the molecule, like for example alkyl chains to improve solubility. Organic semiconductors usually have quite broad gain spectra, especially in the case of macromolecular conjugated polymers, which allow a fine-tuning of their emission without changing material. Lasing from these material generally occur in the visible range. Conjugated polymers with band-gaps in the near-IR generally do not have efficient luminescence emission.
Lasing in organic materials involves radiative recombination of singlet exciton. The lifetime of exciton is a few hundreds of picoseconds to nanoseconds. As in the case of fluorescent dyes and proteins, this short lifetime requires high pump rates to reach laser threshold, and organic semiconductor lasers have been demonstrated only in pulsed regimes. Organic semiconductors have lower charge mobility compared to crystalline inorganic semiconductors, and polarons (electron-phonon coupled charge carriers) have significant absorption, leading to high loss at high carrier densities. For these reasons, electrical pumping of organic semiconductors has not been successful to date.
The stability of organic semiconductors is an important factor to consider for use in biological environment. These materials are usually very sensitive to physico-chemical interactions with oxygen and water, which can be particularly detrimental for both charge transport and fluorescence emission. Nonetheless, organic semiconductors offer several properties that are potentially attractive for laser particles in biological environments. One is generally superior biocompatibility compared to inorganic materials due to their organic compositions, more compliant mechanical properties, and cell-friendly surface roughness. Some organic materials also allow both ionic and electronic conduction, making them suitable to interface with biological tissues where electrical signal are generally conveyed by ion movements. Moreover, the organic materials can be chemically tailored with functional groups sensitive to specific analytes. Chemical and biological sensing based on fluorescence modulation and even laser emission from conjugated materials has been reported.
Ions of rare-earth or transition metals included as impurities in transparent solid-state hosts are well established gain elements for lasers. The first laser demonstrated in 1960 by T. H. Maiman used a ruby which has Cr3+ ions doped in a Al2O3 crystal. The host materials may be an oxide crystal (e.g. Al2O3 and Y3Al5O12 or YAG), fluoride crystal (YLiF4 or YLF), glass (SiO2 or silica), or polymer (polystyrene). Compared to crystalline hosts, amorphous materials can be more easily manufactured into micron sizes, and their optical attenuation and thermal resistance, which are typically higher than crystals, are desirable for small lasers. The most common dopant ions used for gain elements include trivalent lanthanides neodymium (Nd3+), erbium (Er3+) and ytterbium (Yr3+), as well as the transition metal titanium (Ti3+) and chromium (Cr2+, Cr3+, Cr4+). They usually have multiple electronic transitions that can support laser action and their emission properties also depend on the particular host material. In most cases lasing is in the near-IR region between 0.9 and 1.6 μm, but some systems at shorter (like ruby at 694 nm) and longer (like Er:YAG at 2.9 μm) wavelength are also present. A common characteristic of ion-doped solid state lasers is that the electronic transitions from the upper to ground states are much weaker so that the relaxation time, τs, can be quite long, in the order of μs to 10 ms. The long relaxation time is well suited for continuous operation.
Given long lifetimes, rare earths are suited for upconversion to generate emission at higher energy than pump photons. The process involves the absorption of two photons in sequence to promote an electron to metastable state and then an excited state, from which the electron decays to the ground state, emitting a single photon with about twice the pump photon energy. Since the intermediate state has a long lifetime, the process is quite efficient. Upconversion is distinct from typical nonlinear two-photon absorption, where the intermediate state (and also the final upper state) is a virtual state with a lifetime of only a few femtoseconds, and thus efficient excitation can be achieved only with ultrafast pulses with high peak intensity. The detailed photophysical pathways are extensively reviewed elsewhere. The first laser based on upconversion was demonstrated in 1971, and the process attracted significant interest because it allows to obtain visible emission using near-IR pumping, particularly for biomedical applications as luminescent nanoparticles. This material is an attractive candidate for laser particles, with main advantages of near-IR pump light, which has deeper penetration depths and is less harmful than visible light in tissues.
An optical resonator, or laser cavity, provides phase-coherent optical feedback, allowing for regenerative amplification in the gain medium. Various optical elements, including the gain medium, may be incorporated into a laser cavity to control and condition the optical characteristics of intracavity light and the output laser properties.
Cavity modes are particular spatial (transverse) and spectral (longitudinal) optical waves that that can resonantly oscillate in the cavity. Transverse modes are determined by the shape and size of the cavity along the dimension parallel to the axis of optical oscillation. Longitudinal modes are a set of frequency components that satisfy that the total optical phase accumulated in one round trip propagation is equal to an integer multiple of 2π. This resonance condition is expressed as:
where l is a positive integer, is the optical frequency of the l-th longitudinal resonance mode, Lc is the round-trip length of the cavity (e.g. twice the length of a linear cavity or 2π times the radius of a ring cavity), and n is the (effective) refractive index of the l-th mode with a frequency of vl. Since the minimum l is 1, the smallest length scale for a resonator is equal to the free-space wavelength λ divided by the effective refractive index n of the resonator.
High n is beneficial for small lasers as it allows small Lc. Semiconductor materials are therefore attractive materials to realize resonators (as well as gain media). Plasmonic polariton resonators made of metals are another attractive approach that takes an advantage of the fact that the real part of the refractive index is enhanced by the coupling of light with the plasmonic, collective motion of free electrons confined in the cavity.
The number of cavity modes is proportional to the volume of the cavity. As the cavity size becomes comparable to the optical wavelength, the total number of radiation modes decreases, and the spontaneous emission factor β increases (see Eq. 1). As β approaches to 1, the majority of absorbed pump energy goes to the cavity modes. At an extreme case, when the size is less than half the optical wavelength, the number of modes becomes one, a situation known yield a “threshold-less” laser.
The optical cavity can affect the spontaneous emission rate of the gain element in the gain medium through coupling to the cavity modes. The enhancement of spontaneous decay rates of confined modes, called the Purcell effect, is a factor that may be taken into advantage in micro and nano-scale lasers.
When the size is comparable or less than the optical vacuum wavelength λ, the laser behaves similarly to a point source, emitting light over a broad solid angle. Then, the so-called “directionality” of the laser output becomes less pronounced. As the size approaches to the minimum possible size (i.e. Lc=λ/n), the laser output emitted out of the cavity would radiate in all directions. The laser output is emitted in all directions, just like the fluorescence light from a single molecule. However, some additional features such as asymmetry may be incorporated into a laser cavity to enhance directionality.
In terms of the geometrical shape and optical resonance axis, laser cavities can be categorized into various types. Examples are illustrated in
The cavity lifetime τc, a reciprocal of the cavity loss rate, is related to the Q-factor of the cavity, which is defined by:
Q=2πντc, (16)
where ν is optical frequency. Different cavity modes may experience different cavity losses and, therefore, have different Q-factors. In small cavities, the cavity lifetime is typically much shorter than optical lifetime of spontaneous emission, τs. For example, for λ=532 nm, τc=0.28 ns when Q-factor is 106, and τc=280 fs when Q-factor is 103.
From Eq. (8), the threshold pump rate is inversely proportional to the Q-factor:
The spectral linewidth of passive cavity (below threshold) is given by
Above threshold, the linewidth of laser modes can be reduced because the cavity loss is compensated by the gain. According to the Schawlow and Townes limit, the fundamental limit of laser linewidth can be expressed as:
In practice, several factors contribute to broaden the laser linewidth. Thermal heating during laser operation changes the refractive index neff of the cavity, leading to linewidth broadening. In semiconductor lasers, carrier-induced index modulation is an important linewidth broadening factor.
In general, pulsed pumping results in broader spectral widths than continuous wave pumping. The modulation of the temperature and refractive index of the cavity causes the resonance frequencies of lasing modes to be chirped over time, broadening the output spectrum. Relaxation oscillations during the laser build up enhance the modulation. In miniature lasers operated at a pulsed regime, actual laser spectral widths can be several orders of magnitude greater than the theoretical limit in Eq. (18). In these many practical cases, the cavity Q is not a direct deciding factor of the laser linewidth, although it is important to reach lasing threshold (Eq. 17).
Consider a linear or concentric resonator that is formed by a pair flat or spherical mirrors with reflectivity of R1 and R2, and separated by a distance L (Lc/2) The cavity lifetime is expressed as:
where n is the refractive index of the medium in the cavity, and η is fraction internal loss per pass, which can occur due to absorption or scattering loss in the cavity.
Consider a Fabry-Perot resonator (
A photonic crystal cavity has a periodic structure and defects. By having laser gain in the defects and high enough Q-factor, lasing can be achieved. The periodic structure can be either in 1D, 2D or 3D (
Consider a periodic slab of m double layers of two alternating quarter-wave-thick materials, with refractive indices n1 and n2, respectively. The Bragg wavelength λB is given by
λB=2(n1+n2)Λ, (21)
where Λ is the periodicity of the double layer structure. The reflectivity at the Bragg wavelength is: R1,2=(n12m−n22m)2/(n12m+n22m)2. For example, at least 4 double layers of a semiconductor material with n1=3.5 and a low-index interspace with n2=1.33 are needed to achieve 99% reflectivity. Two such stacks on each side of the optical cavity would have a total thickness (8Λ) equal to about 1.6 μm. This would be small enough to be used inside cells.
Whispering-gallery modes (WGMs) are supported in round-shape resonators, such as a sphere, toroid, cylinder, and ring (
WGMs can be designated by the radial mode number q, polar mode number l, azimuthal mode number m, and polarization p. For a sphere the frequencies of WGMs can be approximated (for l>>1) using an asymptotic expansion:
where n1 is the refractive index of the sphere, n2 is the index of the surrounding medium, χ=1 for TE modes and χ=n2/n1 for TM modes, a is the sphere radius, and αq are negative q-th zeroes (−2.3381, −4.0879 . . . ) of the Airy function. The above formula can be further approximated to 2πn1a/λ=l, which is consistent with Eq. (15) where Lc=2πa.
When the size of an WGM resonator approaches to the optical wavelength, its Q-factor is limited by radiative (curvature) loss:
Q≈½(l+½)nr1-2k(nr2−1)1/2e(2l+1)(η−tanhη) (23)
where
is the relative refractive index, and k=0 for TE modes and k=1 for TM modes. For a sphere containing few mM of a typical organic fluorescent dye and nanosecond pumping, the minimum Q-factor required to achieve lasing is ˜10,000. From the above equations, the following approximate formula can be obtained:
For a given Q-factor, high n1 is required to reduce the resonator size. For example, to achieve Q=104 with 2a<1 μm in the cytoplasm (n2=1.37) at λ=0.6 μm, the resonator material should have n1>2.61.
In cavities made of dielectric materials, optical energy is stored in photonic modes, in length scales equal to or greater than the optical wavelength. Better confinement can be achieved with metals by, for example, coating thin metal layers on sub-wavelength dielectric resonators. This configuration can support two distinct types of resonance modes: purely reflective cavity modes and plasmonic-based resonance modes.
In the first case, the intracavity energy is stored in optical waves, just like in dielectric cavities. The metallic layer acts as a mirror that confines the mode inside the cavity by high reflection at the metal surface. The efficient confinement provided by the reflective surface allows high Q factors even for dimensions at the size of optical wavelength.
The second type of modes is based on plasmonic effects that arise at the metal-dielectric interface. The intracavity energy is stored in a plasmonic mode that is strongly confined at the interface between the dielectric medium and metal with a wavelength much shorter than optical wavelength. Therefore, the plasmonic mode resonance can be achieved with cavities much smaller than the optical wavelength. The plasmonic modes are generally divided into long-range surface-plasmon-polaritons (LR-SPP's) and localized surface plasmons (LSP's).
Long-range polaritons are electromagnetic waves that propagate along interfaces between metals and dielectric materials. They are evanescently confined in the direction perpendicular to the surface. Typical structures include metal/insulator/semiconductor/insulator/metal waveguides and metal-insulator/semiconductor nanowires. The plasmonic mode is confined in one or two dimensions mainly in the insulating regions, while the cavity is delimited along the propagation direction by the end facets of the structure, still on the order of (at least) several wavelengths. It should however be noted that the strong coupling of the electromagnetic mode with the electron in the metal leads to higher resistor-type (ohmic) losses that are intrinsic of the plasmonic modes and thus difficult to minimize. Therefore, their Q factors are lower than similar structures based on photonic modes. Lasing has thus far been achieved only at cryogenic temperatures.
Localized plasmons are collective oscillations of electrons at the surface of metallic nanoparticles by coupling with an intense electromagnetic field. The surface plasmons act as resonant electric dipoles, with resonance frequencies determined by the shape of the nanoparticle and the permittivity of the metal and surrounding dielectric medium. Plasmonic lasers have been demonstrated by coating a metallic nanosphere with a dye-doped dielectric shell. Upon optical pumping of the dye, the optical energy is readily transferred to spectrum-matched plasmonic resonance of the metallic core via FRET. The excitation of the surface plasmons produces a strong local field that stimulates an increased coupling of emission from the dyes into the plasmonic modes. This feedback mechanism is referred to as surface plasmon amplification by stimulated emission of radiation, and the device operated based on the mechanism is known as a spaser.
Despite some experimental observations, stable realizations of spaser have been difficult due to a low Q-factor and metallic heating, as well as the limited number of gain elements in the nano-scale volume. For example, a 40-nm sphere can contain only 40 dye molecules even at a high concentration of 2 mM (1.2×1018 cm−3). To date, no convincing evidence for the experimental realization of a spaser at a single particle level has been reported. All experimentally demonstrated spasers have dimensions greater then several μm including substrates.
Conventional cavities described so far offer well-defined optical resonant paths. In the case of random lasers, light is trapped in a gain region in a disordered media by multiple scattering (
Lc≈(ltlg/3)1/2 (25)
Consider, for example, a medium containing TiO2 nanoparticles as scatterers and rhodamine 6G as gain dye. TiO2 nanoparticles with a diameter of 200 nm at a high concentration of 5.6×1010 cm−3 have lt=35 μm. The gain length of rhodamine 6G solution at a high concentration of 10 mM is ˜10 μm. Such a random laser cavity offers a critical size of Lc≈10 μm. That is, when a volume greater than 10×10×10 μm3 is illuminated with a strong pump light, laser emission in principle can be generated. A typical skin tissue has lt=50 μm, and an administration of 10 mM rhodamine 6G into the tissue can give Lc=13 μm.
The emission spectrum of a random laser usually contains several narrowband peaks at irregular wavelengths with different intensities. The spatial profile of the output emission is quite complex, resembling speckle patterns. In principle, the laser spectrum and spatial profile are deterministic from the specific distribution of scatterers. However, light scattering from multiple scatterers is difficult to analyze. Nonetheless, the sensitivity of the laser output to the scattering medium may be useful for sensing.
For a laser to operate and generate light, energy in some form needs to be supplied. The input energy absorbed by the gain medium should be sufficient to produce population inversion in the gain medium and generate sufficient amplification to overcome the loss in the cavity. There are a number of different ways to pump laser particles. Optical pumping offers the most convenient way and well suited for situations where the laser particles are readily accessible with light; for example, for laser particles placed in cells in culture or implanted in a tissue at shallow depths.
Current injection is an established method for driving conventional semiconductor lasers. When semiconductor particles are embedded in biological systems, electrical pumping may be possible but requires a clever way to draw sufficient electrical current and voltage across the semiconductor medium to produce net gain. Wireless electrical energy transfer is one option, which has shown to be able to drive millimeter-sized light emitting diodes (LED's) embedded in animals. As the devices are to be further miniaturized, corresponding antennas get smaller, increasing the resonance frequency required for efficient wireless transfer. When the size approaches to micron or submicron scales, the optimal frequency range approaches to the optical frequency. This makes wireless pumping equivalent to optical pumping. In this section, we describe the fundamental of different pump methods. Furthermore, we discuss the exciting, yet elusive, potential to operate laser particles using bio-energy drawn from living systems.
Optical pumping is a convenient way to provide energy to operate laser particles. Let ε the absorption coefficient of a gain medium. It is related to an absorption cross-section σa: ε=ln(10)−1σaNg/V [see Eq. (3)]. The absorption efficiency ηabs, or the fraction of incident pump energy absorbed by a laser particle, is determined approximately by ε times the diameter, 2a, of the particle:
ηabs=1−10−ε(ν
where νp is the optical frequency of the pump light. For example, consider fluorescent dyes with a molar extinction coefficient of 65,000 M−1 cm−1. A gain medium comprising the dye at a concentration of 1 mM has ε=65 cm−1. A film of the dye with a thickness of 10 μm can absorb 14% of the pump beam (ηabs=0.86).
The absorption of a semiconductor is proportional to the carrier density, ρc. The absorption coefficient of a bulk semiconductors with direct band gap is expressed as ε=A(νp/νg−1)0.5, where A=104-105 cm−1 in most bulk semiconductors and νg is the band gap frequency. For example, ε is 1.5×104 cm−1 for GaAs (hνg=1.42 eV) at νp/νg=1.2. A thin GaAs slab with a thickness of 250 nm can absorb 58% of the pump energy (ηabs=10−1.5*0.25=0.42).
Consider a pump pulse with energy of Ep and a beam size. The absorbed energy is then given by Eabs=ηabsEp*ηurea, where an overlap parameter ηarea is equal to the ratio of the size of a laser particle to the beam size and 1 when the beam size is smaller than the size of the gain medium. Part of the total absorbed energy is radiated out of cavity in the form of light via spontaneous and stimulated emission; the rest is converted to heat and other forms of energy.
As the pump intensity increases, more and more gain elements are excited to the upper states. From Eq. (1) and (3) and Ntot(t)=Ng(t)+N1(t), the fraction of gain elements in the emission-ready upper level in the steady state can be shown to be:
When signal-induced gain saturation (i.e. the last term in the denominator) is neglected (βq<<1, near and below threshold), a 50% inversion (N1=0.5, Ntot=Ng) is reached at pump intensity equal to
The 50% pump intensity corresponds to one photon per absorption cross-section per upper-level lifetime, divided by the quantum yield. The 50% population inversion can be achieved by optical pumping for highly efficient gain medium. For example, the absorption cross-section of eGFP is σa=2.1×10−16 cm2 and QY=0.8 for pump pulses at λ=488 nm (hνp=4.1×10−19 J). When the pulse duration is equal to the lifetime (τs=τp), the pump flux, Ip,50% τp, required to excite a half of molecules upper level is about 2.4 mJ/cm2 or 24 pJ/μm2. This fluence level is lower than the typical safety limits for human skin tissues in terms of photothermal damages, by 1 to 3 orders of magnitude depending on the pulse width.
Chemical energy can be employed to create population inversion. One known example is found in hydrogen fluoride lasers, which utilizes the rotational-vibrational excitation of hydrogen fluoride by an elementary reaction of F+H2→H+HF*. This reaction releases a free energy of ΔG=−31.6 kcal/mol. Part of this chemical energy is converted to photons in a spectral range of 2.7-2.9 μm (9.8-10.6 kcal/mol). Typical exothermal chemical reactions based on rotational-vibrational excitations, however, produce ΔG of only 1-10 kcal/mol. This is much smaller than the energy of visible photons (e.g. 40-70 kcal/mol for at 400-700 nm). As a result, all conventional chemical lasers based on such chemical energy operate in the infrared region.
Chemical reactions that involve electronic, instead of vibrational, transitions may be suitable for pumping visible or near-IR lasers. Electronic transition is relatively rare in chemical reactions, and so-called “chemically initiated electron exchange luminescence” (CIEEL) reactions uniquely support electronic transitions. CIEEL can be categorized into two types. Type-1 is intramolecular CIEEL and utilizes energy transfer in the solvent cage and direct light emission from the reaction product: symbolically A+B→[C+D*] and C+D*→C+D+hν. Type 2 is intermolecular and uses a fluorescent dye to emit light: A+B→C+D*; D*+E→E*+D; and E*→E+hν, where E represents a dye.
The bioluminescent and chemiluminescent reactions are potential candidates for pumping laser particles. The high-energy intermediate state can emit light as electromagnetic radiation (type 1) or transfer the photonic energy to a nearby luminescent particle, such as dye, via near-field FRET (type 2). When FRET occurs from bioluminescent molecules, it is often referred to as ‘bioluminescence resonant energy transfer’ or BRET.
The kinetic condition demands that the pumping reaction rate should be much faster than the decay rate. To achieve a high reaction rate, conventional chemical laser employs gas phase reactions. For instance, rate-determining step of pumping reaction of hydrogen fluoride laser is H+F2→F+HF*. The kinetic constant of this reaction is 1.6×1010 M−1 s−1, similar to the typical kinetic constant of 109 to 1010 M−1 s−1 for diffusion-controlled bimolecular reaction at room temperature. Comparing to this, liquid phase chemiluminescent reactions have lower kinetic rates. For example, the typical kinetic rate constant of luciferase-luciferin is only 10 s−1. The kinetic rate constant can be higher, 105 s−1 for the peroxioxalate-rubrene dye reaction, and 107 s−1 for the luminol-peroxidase reaction with a phenol enhancer. The latter is only 1,000 times slower than the reaction in a hydrogen fluoride laser.
The oxidation of glucose releases a large amount of Gibbs free energy: C6H12O6 (s)+6O2 (g)→6CO2 (g)+6H2O (l), ΔG=−685.7 kcal/mol. This chemical energy can theoretically be used to generate light, for example, via subsequent chemiluminescent reactions. The metabolic energy is consumed to produce adenosine triphosphate (ATP) molecules, which are known as the energy currency of cells. This process involves adding a third phosphate group to adenosine diphosphate (ADP). A total of 30 or 32 ATP molecules are generated from one glucose molecule. At any given time, about 1 gram of ATP is present in our body. When cells require energy, ATP molecules are hydrolyzed to dissociate into ADP and phosphate. On average, in a typical cell, 1-2 million ATP molecules are hydrolyzed each second and make free energy available for various cellular functions, via a process: ATP (aq)+H2O (l)→ADP (aq)+Pi (aq), ΔG=−7.3 kcal/mol. This energy is not enough to generate photons in the visible and near-IR range directly. Some bioluminescent organisms, such as fireflies and quorum-sensing bacteria, require ATP as a co-factor to overcome the activation barrier in the chemiluminescent reaction. Harnessing metabolic energy as pump source for laser particles may be an option.
Possible methods for electrical pumping include electro-chemiluminescence, current injection into semiconductors, and electric discharges in gases. Gas lasers, such as He—Ne laser and CO2 laser, use electric current that is discharged through a gas. Electrons are accelerated by high voltage (300-1400 V), and their collision with gas molecules generating plasma. The excited molecular ions emit light directly or transfer energy to other gain molecules.
The general mechanism of electro-chemiluminescence involve electron transfer reactions at the surface of electrodes and annihilation of radical intermediates: (1) A→A.++e− (oxidation at electrode); (2) B+e−→B.− (reduction at electrode); (3) A.++B.−→A.+B (excited state formation); and (4) A.→A+light. The reaction occurs near the diffusion-controlled limit when ion annihilation is the rate-determining step. The kinetic rate of the ion-annihilation step of 9,10-diphenylanthrascene (DPA) is 2×1010 M−1 s−1. An electro-chemiluminescence pumped DPA laser has been demonstrated by using mirror electrodes with 10 V.
In electrically pumped semiconductor lasers, the p-n junction is forward biased with a voltage (typically at 1.5 to 2.5 V). When forward biased, electrons injected into the n-type side flow to the p-type side where electrons in the conduction band are recombined with holes in the valence band. This radiative recombination process can be described with the following simplified rate equations:
Here, N1 is the number of carriers in the gain region, N0 is the carrier population at optical transparency, q is the number of photons in the cavity, I is the injection current, η is the coupling efficiency (equivalent to quantum yield), e is the electron charge, g is a gain coefficient (≈β/τs). Typical semiconductor materials have N0/V=˜1018 cm−3, so for a micron-size laser with V=10−12 cm3, N0=106. Using η=0.4, τc=1 ps, τs=3 ns, the threshold current in the steady state is given by Ith≈eN0/(ητs)=˜130 μA.
Although electrical pumping is commonly employed in semiconductor lasers, the conventional device architecture may not be directly applicable for stand-alone laser particles. We may consider both wire and wireless approaches. It is relatively straightforward to employ some types of narrow electrical wires to supply the necessary bias voltage and current. Wireless power transfer is more attractive for in vivo applications. A radio frequency transmission delivered an electrical power of 4.1 mW to drive semiconductor LEDs at a distance of 1 m, generating an optical power of 7 mW/mm2. Nevertheless, both wired and wireless methods are unlikely to be feasible for standalone micron-sized laser particles. As the size of the lasers decreases, the efficiency of antenna decreases unless the wavelength of the electrical signal decreases proportionally. This demands frequency in the optical frequency, making the wireless transfer equivalent to optical pumping.
Given this challenge, one may consider the potential of harvesting electrical current and voltage from surrounding biological matters. One relevant example is bio-fuel cells, which convert chemical energy into electricity. For example, generations of 0.5-0.8 V of open circuit potential were achieved in mitochondria-based anode using pyruvate and live clam-based anode using glucose. Electrical eels are capable of generating 600 V. Artificial battery inspired from electric eel may enable electrical pumping of laser particles.
Other types of energy include kinetic energy generated by muscles and body movement, which can be converted to electrical energy by having piezoelectric materials, and sono-luminescence.
When a pump pulse is absorbed, part of its energy is converted to heat and increases the temperature of the gain medium. For pump pulses shorter than <10 ns, heat dissipation to the surrounding aqueous medium is negligible during the pulse duration. In this case, the peak temperature increase ΔT immediately after the absorption of pump energy Eabs is given by:
ρVcVΔT=(1−QY)·Eabs (31)
where ρ is the mass density, V is volume, cV is specific heat capacity. The factor (1−QY) represents the fraction of absorbed energy converted into heat.
Let us consider three examples. Case I: for an oil droplet with a size of 20 μm (3.8 ng) and a specific heat (cp) of 1.8 J g−1 K−1, pump energy of 10 nJ on the droplet (i.e. 3.1 mJ/cm2 in fluence) produces heat energy of 4.5 nJ and causes a temperature increase of ΔT=0.7° C. Case II: for a polystyrene bead with a size of 10 μm (0.52 ng) and cp=1.4 J g−1 K−1, pump energy of 1 nJ on the bead (i.e. 1.3 mJ/cm2 in fluence) generates heat energy of 0.45 nJ and a temperature rise of ΔT=0.6° C. Case III: for a GaAs disk with a diameter of 1 μm and thickness of 250 nm (˜1.1 pg=5.8 kg/m3 (ρ)*0.2×10−12 cm3) and cp=0.33 J g−1 K−1, a pump energy of 5 pJ on the disk (0.6 mJ/cm2 in fluence) generates heat of 2.2 pJ and ΔT=6.1° C.
Excessive temperature increase can cause harmful effects to the surrounding medium. A representative effect is protein denaturation, which induces unfolding of protein structure and leads to the loss of protein activity. The free energy change on unfolding of a typical protein in water is 10-20 kcal mol−1. This corresponds to activation energy Ea of 0.4-0.9 eV per molecule. According to the Arrhenius model, the rate of degradation is proportional to e−E
For pulsed heating, the magnitude of thermal injury, represented by protein denaturation, depends on not only temperature, but also the duration of the exposure to the elevated temperature. When the temperature increases for duration of τe, the amount of denaturation or the magnitude of thermal damage is given by:
If the heating time is reduced, considerably higher temperatures are required for denaturation. For example, protein denaturation occurs at a peak temperature of 90° C. (ΔT=53° C.) when exposed to heat with a short duration of 10 ns, whereas protein denaturation is induced at temperature of 60° C. (ΔT=23° C.) for a long exposure of 1 s.
The internal pressure change due to the temperature rise is given by
p1−p0=αVKB(T1−T∞) (33)
where αV is the coefficient of thermal expansion (2.1×10−4 K−1 for water, ˜4×10−4 K−1 for soft tissues, and ˜1.7×10−5 K−1 for GaAs), KB is (isothermal) bulk modulus (2.2×109 Pa for water, and 8.6×1010 Pa for GaAs), p0 and p1 are pressure before and after the pump pulse. A temperature increase by 1° C. gives rise to 0.4 to 1.5 MPa for incompressible solids and liquids.
The pressure and thermal expansion produces work. Part of this energy gives rise to mechanical (acoustic) waves. The acoustic energy is expressed as:
The efficiency of photoacoustic generation, defined as the ratio of the acoustic energy to the absorbed heat energy, is given by:
Here, Γ=αVKB/(ρcV) is Grüneisen parameter. Its value is ˜0.11 for water and cells, and 0.7-0.8 for soft tissues. The coefficient ΓαV is 2.3×10−5 for water (1.3×10−5 for GaAs). In biomedical applications, T1−T∞ is typically limited to <10° C., so only a small fraction (<10−4) of the absorbed energy is radiated as acoustic waves.
The conversion efficiency is not constant but proportional to the temperature rise. It is interesting to recognize that the acoustic energy is proportional to the square of the absorbed energy, rather than linearly proportional to the input energy itself. This relationship may seem counterintuitive but is due to the fact that the initial pressure increase, given in Eq. (34), actually decreases over time during subsequent (adiabatic) volume expansion (
The heat conduction in radially symmetric geometry is described by:
where k is heat conductivity and {dot over (q)} is the rate of heat generation per unit volume. The blackbody radiation of heat has been neglected.
Case I: Short-Pulsed Pumping:
Let us consider a spherical particle with a radius R that is initially in thermal equilibrium with the medium at T=T∞. The particle is then pumped by a short optical pulse at t<0 and was heated to a temperature of T1 at t=0. For simplicity, we assume that the particle and medium have identical material properties including heat conductivity. Convection, which originates from the motion of molecules in the surrounding medium, is neglected during the short time scale. The boundary condition can be rewritten as:
The Green function of the differential equation is
where α=k/ρcp is the thermal diffusivity of the medium. At a given t, the Green function has a 1/e half-width at r=(4αt)1/2.
Therefore, the time that takes for the heat to diffuse across the sphere with a diameter of 2R becomes:
τ=R2/α (40)
This time τ is called the thermal relaxation time of the sphere particle.
Using the Green function and the above boundary condition, the solution can be shown to be:
where ΔTnor ≡(T(r,t)−T∞)/(T1−T∞) is the normalized temperature profile, and erfc=1−erf denotes the complementary error function. In
When the pulse width is shorter than the heat relaxation time, all the heat accumulated in the particle cannot be diffused out to the medium during the pulse absorption. When the medium is tissue or cells, its thermal properties are similar to those of water: So, we may use k=0.6 Wm−1K−1 and α=0.14×10−6 m2 s−1. The relaxation time τ is the fundamental time constant that governs the cooling of particles by heat dissipation to the medium. The time constant is proportional to the inverse of thermal diffusivity. The aqueous medium, such as cytoplasm and tissues, is a better thermal conductor than the air, but more insulating than semiconductors. So, heat dissipation in the biological medium is usually slower than that in on-chip semiconductor lasers. It should be noted that τ is proportional to R2 or the surface area of the particle. The larger (smaller) the particle, the longer (shorter) the heat relaxation time is; For R=10 μm, τ=700 μs; for R=1 μm, τ=7 μs; and R=100 nm, τ=70 ns.
The analysis above is valid for homogeneous materials and a reasonable approximation for particles that have high water contents or are composed of liquids and polymers, such as oil (k=0.15 Wm−1K−1), polystyrene (k=0.14 Wm−1K−1), and silica (k=1.4 Wm−1K−1).
Laser particles made of semiconductors and metals have thermal conductivities 2-3 orders of magnitude greater than those of dielectrics: for example, GaAs (k=52 Wm−1K−1), silicon (k=150 Wm−1K−1), silver (k=410 Wm−1K−1), and diamond (k=˜1000 Wm−1K−1). In this case, the temperature within the particles is nearly constant. This condition is often described in terms of the Biot number, Bi=hR/k1, where h is the thermal transfer coefficient and k1 is the thermal conductivity of the particle. For a short time scale comparable to the characteristic time, e.g. t<0.2τ, h is time varying but may be approximated to h≈k2/R, where k2 is the thermal conductivity of the medium, and, thus, Bi≈k2/k1. The Biot number corresponds to the ratio of the convection at the surface of the particle to the conduction within the body. Small Biot numbers represent small resistance to heat conduction and small temperature gradients within the particle.
Case II: Continuous Pumping:
We now consider that the pump energy is continuously supplied to a particle at a rate of {dot over (q)}p and calculate the temperature of the particle when it reaches the steady state. From the thermal equation in the medium, i.e. ∇2T=0 and boundary condition of −k(∂T/∂r)r=R={dot over (q)}R/3, the surface temperature T1 of the particle is:
where the energy rate {dot over (q)}p per unit volume is given by {dot over (q)}pV=ηabsPp*ηarea. The temperature rise is proportional to the relaxation time of the medium.
The requirement on optical radiation is quantified in terms of maximum permissible exposure (MPE), which is defined approximately as one tenth of the damage threshold of harmful photothermal and photochemical effects. The ANSI Z136.1 standard classifies lasers according to the biohazard to the retina and skin. According to the standard guideline, the laser exposure limits of the skin are given by 0.02*CA J/cm2 for τp=1-100 ns and 1.1*CA τp0.25 J/cm2 for τp=100 ns to 10 s, where CA is a wavelength-dependent empirical coefficient: CA=1 for a spectral range of 400-700 nm, CA=100.002(λ−700) for 700-1050 nm, and CA=5 for 1050-1400 nm. For long exposure τp>10 s, the MPE level is 2*CA W/cm2, which is given in terms of optical intensity rather than fluence because the thermal equilibrium between laser-induced heating and conductive cooling is reached during the long exposure. For long optical exposure of collimated near-IR light at 750-1050 nm to the eye, the MPE level for retinal safety is CA mW/cm2, where the intensity is measured at the cornea.
Integration of a laser source into biological tissues or living cells requires a standalone and compact device with micrometric dimensions (a typical mammalian cell like a HeLa has a diameter of about 20 μm), a pumping energy source and a good biocompatibility. At the present, there are only a couple of examples of successful integration of lasing structures with a size of typically larger than 8 μm in its largest dimension into cells.
Miniaturization of laser sources has been actively pursued in the past decades with the aim of realizing integrated optoelectronic devices with fast switching times and/or low power consumption. One of the most widespread choices to fabricate small compact lasers is to exploit the whispering gallery modes (WGM) that arise in small structures with a sufficiently high refractive index with respect to the external environment. WGM resonators have been realized in a number of different shapes and materials; the gain medium can be used to make up the cavity itself, or its emission may be coupled to WGMs in a resonator made from an inert material. Polystyrene microspheres doped with dyes have shown to generate lasers upon optical pumping. The smallest microsphere lasers reported to date were about 8 μm, which are limited by the decreasing Q-factor with decreasing sizes. Spontaneous emission from dye-doped microspheres with sizes as small as about 3.5 μm have been reported, but only operation below laser threshold was demonstrated due to the limited cavity Q-factor, gain, and pump energy.
Microdisks are common examples of compact lasers made with semiconductors as the gain media, since these structures can be obtained via lithographic processes from epitaxially grown wafers. They are characterized by lateral dimensions of few to tens of microns, but their thickness can be shrunk down to about 100 nm (corresponding to a limit on the order of λ/4n). Given their high refractive index, which is about 3-3.5 for the most commonly used III-V compounds, the smallest cavities have been demonstrated with inorganic semiconductors. For example, structures with diameter of 2-4 μm, thickness of 100-200 nm and Q-factor up to 12,000-17,000 have been realized for lasing around 900-1000 nm. While much higher quality factors (in excess of 108) can actually be reached with proper engineering of the device shape and surface roughness, such structures have generally considerably bigger diameters (few tens of microns). On the other end, devices as small as ≈600 nm have been demonstrated, albeit with lower Q factors (on the order of 103). However, it is noted that these disk lasers are fabricated on substrates, which have a size much larger than the size of the disks. Certainly, no attempts to coat such disks or to insert disk lasers into cells have been reported.
Several pairs of bioluminescence enzymes and substrates with high transfection efficiency and high brightness in the visible range are known. The spectral bandwidth of bioluminescence is typically 50-100 nm, similar to that of fluorescence. The lifetime of bioluminescence is also similar to typical organic fluorophores, in the order of nanoseconds. The typical bioluminescence reaction rate in biological environments is 0.1-10 photon/s per enzyme molecule. Although this rate is insufficient to generate stimulated emission, bioluminescence can be coupled with optical cavities.
For example, as a preferred embodiment, we have fabricated un-doped polystyrene beads that are coated with luciferases can generate bioluminescence with emission spectra modulated by the cavity modes. Biotinylated Gaussia luciferase (GLuc, 19.9 kDa) was coated on the streptavidin-coated surface of polystyrene micro-beads. The Gluc on the bead's surface oxidizes coelenterazine (CTZ, 423.46 Da) to generate bioluminescence light that is evanescently coupled to the WGMs of the bead (
Stand-alone laser particles, once injected into biological systems such as soft tissues, may be detected and localized by an optical imaging instrument such as fluorescence microscope. However, a novel microscope configuration may be devised to detect the stimulated emission from the laser particles. Consider an excitation beam focused into a medium containing a uniform distribution of probes (
Above threshold, the full-width half-maximum (FWHM) of the PSF, ΔLASE, can be derived from Eq. (6) for axial (z) and lateral (x, y) directions. An approximate form for the resolution enhancement factor is shown to be:
where Δ0 is the diffraction limit of Gaussian beams: Δ0,z=2zR; Δ0,x,y=1.67zRNA, where zR and NA are the Rayleigh length and numerical aperture of the excitation beam, respectively. The resolution is always less than diffraction limit and smallest near the threshold Pth at which (ΔLASE/Δ0)min=β1/4.
LASE microscopy uses single-photon absorption (although two-photon pumping is possible), generates stimulated emission only at the focal volume, and does not require a tight pinhole. In this sense, LASE microscopy has the advantageous features of confocal and two-photon microscopy to achieve efficient generation and collection of signals, with sub-diffraction limit resolution.
Another advantageous feature is low out-of-focus background. The narrow spectral features of the laser output are easily distinguished from auto-fluorescence from other fluorescence molecules that may be present around the laser particle. The decay time of stimulated emission is determined by the cavity lifetime, which is much shorter than fluorescence decay times. Therefore, lifetime measurement can differentiate laser modes from spontaneous background. The low background in LASE microscopy can offer enhanced imaging depths, greater than 500 μm or even 1-2 mm, in scattering tissues.
We constructed an inverted imaging setup (
As laser particles, we used lead iodide perovskite (CH3NH3PbI3) nanowires that were grown to typical widths of 300-500 nm and lengths of 3-7 μm. The samples were transferred to a slide glass and sealed from air with a cover glass optical epoxy to prevent degradation by moisture and oxygen.
Given the long shape of nanowires, the pump beam was shaped to have a matching elongated shape and measured scan profiles across the short axes of nanowires. The pump light source was a microchip laser emitting at 532 nm with a repetition rate of 5 kHz and a pulse duration of ˜2.5 ns, slightly longer than the fluorescence lifetime (τ≈2 ns) of perovskite. The objective lens was a 0.8-NA, 40× water immersion lens (Nikon). A cylindrical lens (CL; f=500 mm) was employed to make an elliptical pump beam profile at the focus, with a FWHM of 2.4 μm along the x-axis and 9.5 μm in the y-axis. The elliptical pump profile was employed to illuminate the whole length of the nanowire and therefore achieve efficient pumping. The nanowire on a sample stage was oriented parallel to the major (y) axis of the pump beam. The polarization axis of the pump beam was aligned orthogonal to the long (y) axis of the nanowire, at which the threshold intensity was the lowest. The pump beam was scanned along the x-axis by translating a collimation lens (L1) resulting to a 35-nm step resolution in the imaging plane. The output emission was directed to the camera and a polarization-independent, diffraction grating-based spectrometer. The entrance slit of the spectrometer was oriented along the long-axis (y) of the nanowire so that the emission from the entire nanowire was collected by the spectrometer. The spectral resolution was ˜1 nm.
The inset of
Experiments were performed to support the proof of concept of LASE imaging.
Similar resolution enhancements by a factor ˜5 were measured consistently from numerous different nanowires, with similar lengths and β values. The opening width of the confocal slit in front of the spectrometer affected resolution only modestly, through its influence on the laser output measurement leading to slightly different β values.
A LASE microscope may be extended from the setup in
A LASE microscope may be further combined with other imaging modalities, such as confocal fluorescence imaging, multiphoton imaging, reflection imaging, and optical coherence tomography (OCT). For example, the system can employ a femtosecond laser to generate and collect two-photon or three-photon excited fluorescence or harmonic generation signals from the structures, such as cells and tissues, surrounding the laser particles. An exemplary embodiment is illustrated in
Instead of optical detection, laser particles may be detected and imaged by using photoacoustic microscopy. Laser particles can release a small fraction of absorbed pump energy as acoustic wave energy. This mechanism may allow the particles to be detected by acoustic transducers as in photoacoustic detection. The amplitude of the acoustic wave is proportional to the pump rate P(t). In an unsaturated regime that Ng≈Ntot, the photoacoustic signal is simply proportional to the pump intensity.
However, the number of molecules in the excited state N1 becomes comparable to Ntot, the depletion of the ground state limits pump absorption and, as a result, attenuates the photoacoustic signal compared to the unsaturated case for the same intensity. It is noted that above threshold, the laser oscillation clamps the number of excited molecules to a constant level because the stimulate emission brings down the excited-state molecules to the ground state. As the gain molecules are recycled more rapidly in lasers, this process can result in higher photoacoustic signals in laser particles compared to the same gain medium without cavity in the saturated regime.
Guide star is an astronomical technique to estimate wavefront distortion due to atmospheric turbulence. Since density of air becomes varied when turbulence exists, heterogeneity of refractive index deforms a wavefront of light, then telescope in ground observatory detects severely degrade star images. To solve this problem, a laser beam is illuminated directly to the sky and wavefront distortion is measured by comparing the difference between incoming light and returning light. The return light can be generated by Rayleigh scattering of laser light from the air molecules at 10 to 15 km altitude or laser-induced fluorescence of atomic sodium at 95 km altitude. After guide star-assisted wavefront measurement, blurred star image can be recovered via an adaptive optics device, such as an array of deformable mirrors to compensate wavefront distortion.
Using same principle in astronomy, guide star also plays an important role in biological imaging. Biological tissue causes scattering and distortion of wavefront due to its inhomogeneous optical properties, limiting imaging of specimen, especially placed at depths. Since transmission matrix provides the estimation of output field after scattering media at given arbitrary input field, it is possible to make a focused spot or correct degrade images of specimen in deep tissue by shaping wavefront of input light based on the information of transmission matrix. In this process, guide star is helpful to measure a transmission matrix of given tissue. The principle for guide star involves feedback and conjugation processes. First, a feedback method utilizes fluorescent particle placed in deep tissue and monitors fluorescence intensity and find out the maximum intensity. Second, a conjugation method collects light emitted from a target and synthetize a phase-conjugated beam that, when illuminated to the tissue, can trace back to the focus. This process is possible because scattering and propagation of light in tissue is time-reversible.
Rather than normal fluorescent label, laser particle can be a potential candidate that can show better performance as a guide star. For example, lasing threshold of laser particle can be a sensitive indicator in feedback guide star. Laser particles can be an ideal source for conjugation guide stars, since they can generate spatially and temporally coherent light. Combining with injection locking that gives temporal phase to emitted light from laser particle, laser particle can be implemented into various applications such as conjugation-based correction using coherence gating.
The coherency of the output can be extended to the optical phase level by using injection locking (
Laser particles may be incorporated into the cytoplasm of live cells. A variety of biological and cellular uptake processes exist, which includes endocytosis, phagocytosis, and macro pinocytosis. Cells may be incubated with a media containing laser particles for sufficient time so that cellular uptake of the particles occur.
For cells that do not uptake particles efficiently, physical insertion methods, such as robotic pipette-based injection, may be used.
More controlled uptake or laser loading process can be achieved by using a microfluidic droplet device. A couple of techniques to generate aqueous droplets encapsulating micro and nanoscale objects in oil using syringe pumps are well known in the art.
Employing a passive or active sorting arrangement, such as pillars and high-voltage electrodes, the apparatus may be operated to sort laser particles or cells carrying or containing laser particles by using the optical information based on the output emission spectra from the laser particles.
Due to their narrow emission spectrum, lasers and optical cavities embedded in cells have great potential for applications in optical tagging of single cells. Cell tagging would enable to distinguish between large number of subpopulations of cells or even studying cells at singe-cell level and for highly multiplexed assays. Current tagging methods include graphical encoding with emitters arranged in specific patterns, physical encoding, lifetime encoding, spectrometric encoding, and mass-based tagging using rare earth elements. Most of these methods are limited to approximately 10 to 100 barcodes. Graphical barcodes enable a larger number of combinations, but they require imaging, which is slow. Spectrometric methods including Brainbow and Multibow are limited by the broad emission of the dyes, quantum dots or fluorescent proteins. The number of unique tags can be increased by multiplexing with intensity and encoding with up to 500 combinations is commercially available. Intensity encoding however requires calibrated instruments and can change due to photobleaching, dye degradation, light scattering and absorption, which could be especially problematic for in vivo studies.
Conventional fluorescence microscopes and fluorescence-activated cell sorters (FACS) cannot resolve fine spectral lines. For high spectral measurement resolution, a flow cytometer apparatus can be configured by employing a pump source and a high-resolution spectrometer (
Laser particles with submicron sizes are coated with binding molecules, such as antibodies, peptides, and click molecules. These laser probes are mixed with cells so that the probes bind with target specific biomarkers on the cellular surface or in the cytoplasm (
To realize miniature laser particles, in certain implementations, photonic semiconductor whispering-gallery-mode disk lasers may be used. Because the refractive index (n=3-3.5) of semiconductors is much higher than that of the cytosol (n=1.33-1.4), semiconductor disk lasers can have a diameter of smaller than the vacuum optical wavelength and a thickness of a fraction of wavelength. Like quantum dots, semiconductor particles do not photo-bleach and biodegrade, making them suitable for long-term tracking.
Millions of identical or incremental particles can be fabricated from a single batch by standard electron beam lithography (
In example implementations, we have fabricated semiconductor micro-disks with diameters ranging from 0.5 to 3 μm. These micro-disks can have an InAlGaAs quantum well structure built on an InP substrate. The disk shape can first be lithographically carved using electron beam lithography, reactive ion etching, and then disconnected from the substrate by acidic wet etching that removes InP sacrificial layers (
When optically pumped at 1064 nm with sufficient pulse energy (1-100 pJ), almost all the laser particles harvested reached laser thresholds and generated narrowband stimulated emission (
The laser output from a microdisk particle can be measured through a relatively thick biological tissue, such as the murine ear skin, with a high signal to noise ratio (
As described in Eq. (22), the output wavelength of laser disks can be tuned by varying the diameter (
Besides diameters, several other variables can be used to tune the output wavelengths. These include different semiconductor materials for the active and clad (shell) regions, strain of the semiconductor materials, the thickness of quantum wells, distance between quantum wells in the case of multiple quantum wells, and the thickness of passivation layers. A grating structure can be formed on the surface of laser particles for mode selection and to achieve single mode oscillation.
Material systems for short wavelengths in the UV, blue, and gree regions are II-VI compounds and III-V nitride systems, such as ZnS, ZnSe, GaN, and AlN. AlGaInP systems, such as GayIn1-yP quantum wells sandwiched in (AlxGa1-x)0.51In0.49P separate confinement barrier layers, can generate visible wavelengths in 600-700 nm. AlxGa1-xAs grown on GaAs lattice are suited to obtain lasing wavelengths in the range of 700-900 nm for double heterostructures. The wavelength region of 900-1100 nm can be converted by the strained InxGa1-xAs sandwiched between AlGaAs or GaAs layers. The range of 1.1 to 1.65 μm is covered by In1-xGaxAsyP1-y grown on InP or AlxGayIn1-x-yAs—InP system. Other systems include InGaAsN/GaAs quantum well structures.
For another experimental demonstration of laser particles and their intracellular operation, we fabricated microdisk lasers from various semiconductor epitaxial wafers. The wafers include InP/InAlGaAs/InP with different alloy compositions for the InAlGaAs gain layers with a thickness of 200 nm (or 350 nm). The nominal gain center wavelengths of the wafers were 1200, 1275, 1350, and 1425 nm, respectively. We also fabricated microdisk lasers from GaAs/AlGaAs/GaAs (840 nm), and InP/InAlAs/InP (840 nm), as well as quantum well wafers.
Electron-beam lithography has high resolution at 10-nm scales but requires relatively long scan times. For rapid fabrication in large volume, we used optical lithography using a custom-made mask and photoresist. After UV illumination and reactive ion etching, the size of disk columns ranged from 1.5 to 3 μm. The patterned substrate was transferred to an HCl solution for wet etching to remove InP sacrifice layers. After filtering and washing, we obtained stand-alone laser particles in colloidal solution. The yield of the wet-etching/transfer process could be as good as 50% by carefully minimizing the loss of particles. Out of 10 million microdisks on substrate, more than 5 million stand-alone laser particles were collected.
Semiconductor materials exposed to the aqueous and biological environment may generate electrochemical effects that can be toxic. At surface, the bond structure is broken. The broken bond makes the surface highly reactive, resulting in surface reconstruction (surface atoms bond among themselves) and surface contamination by the adsorption of O, C, and other atoms and molecules. This can alter the energy levels of semiconductor at the surface, which can result in the leakage of carrier current through the surface to the surrounding environment and generate a number of undesired electrochemical processes such as semiconductor corrosion and water splitting, as well as the degradation of quantum efficiency of the laser particles.
A practical solution to this problem is coating of semiconductor surface with passivating layer so that bulk bonding prevails at the semiconductor surface. The widely used capping materials include SiO2 (
Surface coating or surface functionalization with biocompatible polymers, as used in biocompatible quantum dots, can further increase the biocompatibility of laser particles and make them usable in cells and animals for long terms experiments. Laser particles can be coated with such polymers and biocompatible or bio-inert materials (
Furthermore, laser particles may be encapsulated by liposomes to improve biocompatibility and circulation time (
Genetically tagged probes are useful, for which we may use hybrid approaches. The target cell group in an animal can be made to over-express certain surface membrane proteins, and the laser particles are coated with their antibodies so that they bind to the cell surface. Second, cells are transfected to produce specific intracellular proteins, and particles are designed to produce a certain emission characteristic only upon binding with the proteins in the target cells.
For biocompatible coating, after the transfer of disks off wafer by immersing in HCl and filtering to remove HCl, we mixed disks (1 million/ml in EtOH) with 0.1 μM MPTMS (3-mercaptopropyl trimethoxysilane) to deposit monolayer of silane on semiconductor surface. Then, NH3, H2O and TEOS were added to the solution to grow silica shell. The reaction time and TEOS concentration were adjusted to obtain desired silica shell thickness.
Cells were incubated with isolated laser disk particles to internalize them into the cytoplasm.
Referring to
Consider two laser particles attached to form a doublet. There are 499,500 (=C(1000,2)=D(1)=S(2)) combinations of doublets. There are 166 million types of triplet lasers (formed by three lasers). These strategies provide virtually an infinite number of labels. A key factor for this massive wavelength-division multiplexing is that N is large. With fluorescence, N is limited to four or 10 at most.
Laser-based tagging allows cell lineage tracing. Initially, cells are loaded each with many particles or large multiplicity of infection, for example, <m>=128 of doublet particles with no overlaps. This allows about 3,500 (D(1)/128) cells to be labeled. Now, these cells divide over time. Since the group of particles is split roughly half and half at each division, after 5th generation there will be on average 4 doublets remaining in a cell. Since the same doublets were only present in their direct ancestors, but not others, it is possible to identify their family tree. By counting the number of particles, the approximate number of passage can also be estimated. The map of proliferation can also be measured.
Doublets and triplets can be formed, taking advantage of the discoidal shape (
Alternatively, two laser disks may be coated with two different binding chemicals, such as Tz and TCO click chemistry molecules, respectively, and glued with each other. Another method is to use spacer materials such as polystyrene beads with diameters of 0.3-1 μm. With appropriate chemical glues, doublet (
Besides cell tagging, the ability to observe their proliferation, migrations, and cell-cell and cell-tissue interactions over time in live animals by tracing individual cells would be extremely useful. This is the major advantage over viral labeling techniques, known as DNA barcodes, which cannot be visualized by optical imaging and can be read in vitro only after sacrifice of animal using PCR amplification and sequencing. The laser-tagged cells can be imaged repeatedly in vivo, analyzed by flow cytometry, and sorted for gene profiling and single-cell RNA sequencing. The ability to obtain such comprehensive information from molecular, cellular, tissue, and systems levels over millions to billions of cells in a single animal experiment is unprecedented. It should be noted that the laser probes could be used in conjunction with conventional fluorescence labeling techniques and transgenic reporter mice, as well as DNA barcoding.
Another application area of laser particles is biosensing. Different laser cavities are sensitive to the change in the refractive index change in the surroundings of the cavity, most notably WGM cavities and photonic crystal cavities. Either of these cavities can be operated in lasing regime or just as notch filter to measure the shift in the resonant frequencies. Instead of just measuring the refractive index the surface of the cavity can be functionalized to bind just specific molecules to the surface, in this way specific detection is possible. The WGMs can interact with the environment outside the microsphere through evanescent field. A particle on the surface or near the surface of a microsphere changes the optical path of the light and/or the cavity loss. This causes a shift of the resonant frequencies. The resonating condition in whispering-gallery mode resonator is approximately 2πan1=lλ. Here, a is the radius of the resonator, n1 the refractive index, l the mode number, and λ the resonant wavelength of the mode. The attachment of molecules to the surface increases the effective refractive index.
The index change alters the resonance wavelength via:
where αex is the excess polarizability of the binding molecules, σ is the surface density of the molecules, ε0 is vacuum perimittivity, n1 and n2 are the refractive indices of the resonator and surrounding medium, respectively. Since the detection of mode shift will be limited by fundamental linewidth of the mode, the smallest surface density can be estimated by using Q≈Δλ/λ. For example, bovine serum albumin (BSA) has αex/4πε0=3.85×10−21 cm3, and using a 15-μm polystyrene bead (n1=1.59) in water (n2=1.33) with Q=2,000, a surface density as low as 5.9×1012 molecules cm−2 can be measured, which corresponds to 107 molecules on the surface of the bead.
For example, the Stokes radius of BSA is 3.48 nm, so the effective binding area on the surface of resonator is approximately 3.8×10−13 cm2. If BSA molecules bind on the surface of 15-μm polystyrene bead without any gap, the total number of BSA is 1.9×107. This can be readily detectable with Q=1,000, and this level of sensitivity is comparable with other label-free methods using surface plasmon resonance.
Spherical microresonators with very high Q-factors are among the most sensitive optical systems. In the case of a microresonator with Q=108 and diameter 100 μm the light travels few tens of meters around the sphere and the particle on the surface is sampled more than 100,000 times.
Lasing modes are highly dependent on the geometry of the laser cavity. This can be exploited by making a soft laser cavity and using it as force sensor. Measurement of small forces inside a cell was demonstrated using oil droplet WGM intracellular lasers. When a droplet is subject to uniaxial stress, its shape deforms which is manifested in the emission spectrum as splitting of laser lines. For small forces the shape can be approximated as a spheroid. Lasing is confined to the equatorial plane, which has the lowest optical loss due to minimum curvature. By fitting to a model for non-spherical WGMs, the equatorial and polar semi-axes are calculated giving the average diameter and eccentricity. Flattening stress Δσ is related to the local mean curvature of droplet surface by Δσ=2γΔH, where γ is surface tension and ΔH is the difference in the curvature. For small eccentricity (ε2<<1) the stress is approximated as
For example, the lateral stress exerted onto an oil droplet with a diameter of 8 μm in a HeLa cell was measured to be Δσ=500 pN/μm2 (500 Pa). Due to cellular dynamics the stress changes with time. The mean fluctuation of the internal stress was measured to be ˜150 pN/μm2 (Pa). The sensitivity limit of this method determined from the fluctuations in a dead cell is ˜20 pN/μm2 (20 Pa), approximately an order of magnitude better than direct image-based analysis.
The force field inside a cell and possibly tissue can also be measured using solid polystyrene beads, however due to their high Young's modulus, the minimum force that can be measured is much higher than the case of droplets. Force was measured during cell uptake of a bead with the minimum measured stress in the order of 100 kPa. The mode splitting due to deformation of the cavity is difficult to separate from the splitting due to local change in the refractive index during bead uptake by the cell.
The sensitivity of resonance peaks to temperature can be estimated from the material properties.
where αT is the coefficient of thermal linear expansion and βT is the coefficient of relative index change of the cavity material. For polystyrene, αT≈7.5×10−5° C.−1 and βT=−8.2×10−5° C.−1. Therefore, these two effects almost cancel out each other. Indeed, our measurements show a mode shift of only 3 pm/° C. for a bead immersed in pure water. This corresponds to a diameter error of 60 pm/° C. for a 10 μm bead. This temperature-dependent effect is well within the diameter interval of 2 nm. We note that in mammalian cells in culture or in vivo, the ambient temperature is kept constant at 37° C. within few degrees.
For semiconductors, the contributions of thermal expansion and index change have the same sign. For example, GaAs and InP at 37° C. have αT of 0.59×10−5° C.−1 and 0.46×10−5° C.−1, respectively, and PT at 1.15 μm of 7.8×10−5° C.−1 and 7.2×10−5° C.−1. This yields to a wavelength shift of 60-90 pm/° C. for GaAs and InP lasers.
It is noted that the terms “substantially” (as in substantially different from or substantially the same as), “about,” “approximately,” etc. relative to specified values may indicate appreciably (or not appreciably) different from, within acceptable manufacturing tolerances, and/or without deviation that would significantly impact intended operational parameters. In certain implementations, acceptable values (that are substantially the same as, substantially different from, about, or approximately a specified value) may have, for example, a +/−1 percent deviation, a +/−5 percent deviation, or a +/−10 percent deviation from the specified value, depending on the specific applications. Other acceptable deviations include, for example, at most 1 percent or at most 5 percent (in the case of substantially the same as, about, or approximately), or at least 5 percent or at least 10 percent (in the case of substantially larger/smaller than).
The present invention has been described in terms of example embodiments, and it should be appreciated that many equivalents, alternatives, variations, additions, and modifications, aside from those expressly stated, and apart from combining the different features of the foregoing versions in varying ways, can be made and are within the scope of the invention. While the above detailed description has shown, described, and pointed out novel features as applied to various embodiments, it will be understood that various omissions, substitutions, and changes in the form and details of the devices or algorithms illustrated can be made without departing from the spirit of the disclosure. As will be recognized, certain embodiments of the disclosures described herein can be embodied within a form that does not provide all of the features and benefits set forth herein, as some features can be used or practiced separately from others. The scope of certain disclosures disclosed herein is indicated by the appended claims rather than by the foregoing description. All changes which come within the meaning and range of equivalency of the claims are to be embraced within their scope.
Patent | Priority | Assignee | Title |
Patent | Priority | Assignee | Title |
5248772, | Jan 29 1992 | COULTER INTERNATIONAL CORP | Formation of colloidal metal dispersions using aminodextrans as reductants and protective agents |
9502227, | Nov 02 2007 | HUMANIX CO., LTD. | Capturing of cell fluid and analysis of its components under observation of cells and instruments for the cell fluid capturing and the analysis |
20060046312, | |||
20060123900, | |||
20080159351, | |||
20080204709, | |||
20080241262, | |||
20090114273, | |||
20110253909, | |||
20110256577, | |||
20110266470, | |||
20120013909, | |||
20130009119, | |||
20150076408, |
Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
Apr 12 2019 | YUN, SEOK HYUN | The General Hospital Corporation | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 059059 | /0576 | |
Jun 08 2020 | The General Hospital Corporation | (assignment on the face of the patent) | / |
Date | Maintenance Fee Events |
Jun 08 2020 | BIG: Entity status set to Undiscounted (note the period is included in the code). |
Jun 16 2020 | SMAL: Entity status set to Small. |
Date | Maintenance Schedule |
Mar 29 2025 | 4 years fee payment window open |
Sep 29 2025 | 6 months grace period start (w surcharge) |
Mar 29 2026 | patent expiry (for year 4) |
Mar 29 2028 | 2 years to revive unintentionally abandoned end. (for year 4) |
Mar 29 2029 | 8 years fee payment window open |
Sep 29 2029 | 6 months grace period start (w surcharge) |
Mar 29 2030 | patent expiry (for year 8) |
Mar 29 2032 | 2 years to revive unintentionally abandoned end. (for year 8) |
Mar 29 2033 | 12 years fee payment window open |
Sep 29 2033 | 6 months grace period start (w surcharge) |
Mar 29 2034 | patent expiry (for year 12) |
Mar 29 2036 | 2 years to revive unintentionally abandoned end. (for year 12) |